Age-Related Changes in the Pattern Electroretinogram of Normal and Glatiramer Acetate–Immunized Rats

Abstract

Purpose.: Age-related changes in retinal ganglion cell (RGC) activity may impact results of long-term functional studies of disease progression and drug efficacy in humans and animal models. Though these changes can be evaluated using the pattern electroretinogram (PERG), longitudinal studies suffer from failure of follow-up from birth to senescence. Our aim was to perform a long-term, longitudinal study evaluating age-related changes in the rat PERG, by conducting repeated, serial recordings in the same animals. Additionally, we tested the hypothesis that neuroprotective treatment using glatiramer acetate (COP-1) immunization may delay age-related decline in function.

Methods.: PERG was recorded from six untreated and seven Cop-1–immunized Lewis rats. Recordings were conducted at 2- to 4-week intervals from age 5 to 59 weeks.

Results.: PERG amplitudes significantly increased between 5 to 7 weeks of age, and decreased from age 30 weeks onward (P = 0.016 and 0.0002, respectively). Amplitudes fluctuated insignificantly in weeks 7 to 30, with peak amplitudes reached at age 18 weeks in most spatial frequencies tested. N2 implicit times were shortened, mainly during weeks 5 to 18 and 40 to 59 (P < 0.001). PERG amplitudes of Cop-1–treated rats were similar to controls (P = 0.137) but peaked later (22–26 weeks).

Conclusions.: This 14-month-long study provided accurate measurement of developmental and aging changes of rat RGC function using repeated testing of individual animals. We found functional development to extend beyond the reported period of structural changes. Cop-1–immunized rats were not protected against age-related decline in inner retinal function, although their PERG maturation dynamics were altered.

Introduction

Age-related structural and functional changes in the neural retina are biphasic, as an initial maturation stage is followed by a gradual aging one. Functional maturation of the outer retina in humans, as evidenced by changes in the scotopic and photopic b-waves of the full-field electroretinogram (FERG), is characterized by a decrease in implicit times and an increase in amplitudes during the first year of life, ¹ with responses reaching adult levels by age 3 to 5 years. ² Functional parameters that reflect maturation of the inner retina include a decrease of the implicit time and increased spatial and temporal resolutions of the pattern electroretinogram (PERG) in infants 5 to 20 weeks old. ³ In contrast to the FERG, there is not much PERG data from subjects 1 to 5 years old; therefore, the age at which maturation is achieved is not clear. 

Aging changes of retinal function in humans show gradual deterioration of many ERG parameters. Among those changes are decreased amplitudes and increased implicit times of the FERG a- and b-waves ⁴ and multifocal oscillatory potentials, ⁵ and attenuation of the multifocal ERG ⁶ and PERG. ⁷ A trend of attenuation of the FERG parameters, ⁴ as well as the P50 component of the PERG, ⁸ begins as early as 7 to 18 years of age. 

As long-term, longitudinal studies may be required to study chronic diseases such as glaucoma, ^9–11 it is important to know the age-dependent changes of normal PERG parameters. Rats and mice are used as animal models for studying retinal physiology and diseases. The PERG is often used to evaluate inner retinal function in these models, having been used to study the role of antioxidants, ¹² optic nerve transection, ^13,14 ocular hypertension, and ischemic conditions ^11,15–19 affecting retinal ganglion cell (RGC) survival and function. ^10,20,21 Previously, we have used PERG to evaluate inner retinal function in ocular hypertension models ^21,22 and to describe functional maturation of the inner retina in young rats. ²³ Our goal in the present investigation was to conduct a long-term, longitudinal study of PERG recordings in rats, with the aim of characterizing inner retinal functional maturation and aging changes and describing the transition profile from the former to the latter. 

In addition, we wanted to evaluate the effect of neuroprotective treatment on age-related changes in inner retinal function of the rat. Because decline in RGC density has been proposed to be a factor in age-related changes in human PERG ⁷ and because a similar density decline has been shown in aging rats, ^24,25 we hypothesized that amelioration of this decline using neuroprotective treatment will affect aging changes in the rat PERG. Autoimmunization with glatiramer acetate (Cop-1) was previously shown to reduce RGC loss and preserve inner retinal function in optic nerve crush and ocular hypertension models in rats ^22,26 through controlled activation of T-cells. Inner retinal function of Cop-1–immunized rats was compared with age-matched rats in sequential recordings from age 5 to 59 weeks. 

Methods

Animals

Six male and seven female inbred Lewis rats from two litters were raised from birth to age 59 weeks and were repeatedly tested throughout the experimental period. Six animals served as vehicle-treated controls, and the remaining seven were immunized with Cop-1 (see below). Rats were reared in a 12-hour cycle of dark/light (average illumination of 6 foot-candles [ft-c]) and controlled temperature. Animals were individually tagged, so that age-related changes in individual animals could be assessed. 

This study was approved by the Institutional Animal Care and Use Committee and was conducted in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Visual Research. 

Cop-1 Immunization Protocol

Rats were immunized at age 8 days, followed by booster immunizations on days 30, 48, and 80. Immunizations were composed of 100 μg of Cop-1²² (in 0.1 mL PBS) emulsified with an equal volume of complete Freund's adjuvant containing 5 mg/mL H37 RA (Difco Laboratories, Detroit, MI). Total volume administered was 0.2 mL intramuscular (IM). Controls received an equal volume of vehicle only IM. 

Electroretinography

In each rat, PERG was recorded at ages 5, 7, 9, 11, 14, 18, 22, 26, 30, 34, 40, 44, and 59 weeks. Both eyes were sequentially recorded in each rat, and the first eye to be recorded was randomly chosen. The absence of any anterior segment opacities or evidence of intraocular diseases was verified at the beginning of each recording. 

ERG recordings were performed under general anesthesia (ketamine 70 mg/kg, xylazine 2.5 mg/kg, IM). Recordings followed a previously described protocol. ²³ The active electrode was a contact lens electrode designed for use in rats (Medical Workshops, Groningen, The Netherlands). Reference and ground needle electrodes were placed subcutaneously in the temporal cantus of the ipsilateral eye and over the scruff, respectively. Pupils were dilated with 0.5% tropicamide eye drops (Mydramide; Fischer Pharmaceutical Laboratories, Tel Aviv, Israel). Because visual stimuli were projected directly on the retina (see below), loss of depth of field by mydriasis was of no significance. Lubrication and proper electrical conductance of the active electrode were maintained with 1.4% hydroxymethylcellulose (Celluspan; Fischer Pharmaceutical Laboratories). 

Rats were dark adapted for 10 to 30 minutes only as the PERG is mainly a cone-driven (i.e., photopic) response. ^20,27 Thus, dark adaptation was used only to standardize recording conditions among rats. Visual stimuli consisted of dark and light checkerboards with a 50% duty cycle that alternated at a frequency of 6.1 reversals per second (Viking Quest; Nicolet Biomedical Inc., Madison, WI). Five spatial frequencies of progressively increasing checkerboard size (0.368, 0.184, 0.092, 0.046, and 0.023 cyc/deg) were sequentially tested. Animals were kept in the dark for 1 minute between each stimulus. The mean luminance of the projected display was 100 cd/m², and contrast was maintained at 99%. The stimuli were presented to the animals using a modified table-mounted Bausch and Lomb direct ophthalmoscope (71-63-40; Bausch & Lomb, Rochester, NY), which was fitted with a third optical channel that was used to project the stimulus directly onto the eye. ²⁸ A nodal point was in the plane of the pupil, and the double-beam system used Maxwellian view. A liquid crystal image plate was parfocal with the fundus image and was computer driven. The ophthalmoscope was positioned 7 cm from the eye and aligned with the axis of the globe. Manual control was used to ensure that the stimulus was centered on the eye and focused. Sweep time was 200 ms. Signals were amplified ×500,000 with a 2–250 Hz bandpass (no notch filter was used), and digitized at a rate of 5000 Hz. The response to 100 reversals was averaged. 

When noise levels appeared to be more than 20% of the amplitude, further smoothing of the wave was performed offline using computerized software (Igor Pro ver. 5.0; WaveMetrics, Inc., Lake Oswego, OR). 

Statistics

Paired Student's t-test and ANOVA of repeated measures (every single rat was repeatedly recorded at 13 time points) were used as indicated in the Results section. In this analysis missing data (four instances of single recordings that were discarded due to technical faults) were corrected to the group average at the specific time point. The age-related maturation and decline phases of the PERG amplitudes were fit to four parametric logistic curves (Prism 4; GraphPad Software, Inc., La Jolla, CA). 

Results

PERG Amplitude as a Function of Spatial Frequency

Representative traces from a single session in a 34-week-old control rat are shown in Figure 1. A progressive increase in N2 amplitudes, inversely related to stimulus spatial frequency, is apparent, as was previously documented in the mouse (Porciatti V, Falsini B. IOVS 2003;44:ARVO E-Abstract 2795), rat, ^14,23 cat, ²⁹ dog, ^30,31 and humans. ³² Figure 2 shows the average PERG amplitude of the control rats as a function of spatial frequency at five time points, including start, peak, and end points. At all ages, a saturated response to the lowest spatial frequencies is apparent, as would be expected from the tuning curve of our visual stimuli settings, which consisted of low temporal frequency, pattern reversal, and high luminance (see Bach and Hoffmann for review ³³ ). 

Figure 1. 

 

Representative pattern ERG traces recorded from a 34-week-old control rat in response to decreasing spatial frequencies (top to bottom). Each trace is an average of 100 reversals. Spatial frequency is halved between traces. N2 peak to trough measurement is depicted. Stimuli were projected on the retina using a modified direct ophthalmoscope, at a frequency of 6.1 reversals per second (see text for details).

Figure 2. 

 

PERG amplitude (mean + SD) of control rats as a function of spatial frequency at five time points. A typical saturated response is apparent at all ages. n = 12 eyes. (The x-axis is log scaled.)

Age-Related PERG Amplitude Changes in Control Rats

The age-related dynamics of the PERG amplitudes of the control group followed a course of fast age-dependent increase, followed by a relatively steady mature phase, and finally a slow decline, reflecting maturation and subsequent aging of inner retinal function (Fig. 3). 

Figure 3. 

 

(A) Representative PERG traces from the left eye of a control rat at ages 5, 18, 30, and 59 weeks. Note prolonged implicit time of the N2 wave at 5 weeks, compared with 59 weeks, despite the similar amplitudes. All traces are in response to a 0.023 cyc/deg stimulus. (B) Age-related changes in N2 PERG amplitudes of control rats (+SD, n = 12 eyes of six rats), in response to five spatial frequencies. By 7 weeks of age, mature levels (defined as 75% of maximal response) are reached. A period of insignificant fluctuations that spans ages 7 to 30 weeks follows, and finally a gradual, significant decrease indicating aging changes of inner retinal function. Peak average amplitudes were reached at 18 weeks of age for the 4 lowest spatial frequencies, and at 7 weeks for the highest spatial frequency.

Mature (defined as 75% of maximal response) PERG amplitude levels were reached by age 7 weeks. At ages 7 to 30 weeks, average amplitudes fluctuated insignificantly between 75% and 100% of mature levels. 

Peak amplitudes were reached at age 18 weeks in all but the highest spatial frequency tested. Multiple pairwise comparison analysis with adjustment of the significance level according to Bonferroni criterion (i.e., P < 0.05/12 = 0.0042) revealed that PERG amplitudes at this age were significantly higher than at the 5- and 34 to 59–week time points. Using ANOVA with repeated measures test, for 0.023 cyc/deg the increase in amplitude in weeks 5 to 7 and the decrease in weeks 30 to 59 was found to be significant (P = 0.016 and 0.0002, respectively). The fluctuations during weeks 7 to 30 were insignificant (P = 0.068), thus indicating a stable level period. Therefore, it would be proper to consider the ages 7 to 30 weeks as a mature period that is followed by a period of age-related decline (Fig. 3B, from age 30 weeks onward). A four-parameter logistic curve was fitted to the data of the rising phase (weeks 5–18) (Fig. 4A, r ² = 0.19). The model revealed that the age of half-maximal response (Age50) was 4.4 weeks. Noise level PERG amplitudes at birth were assumed for this modeling. 

Figure 4. 

 

Four-parameter logistic curves fitted to the rising (A) and falling (B) phases of PERG amplitudes (±SEM) of control and Cop-1–treated rats in response to 0.046 cyc/deg stimulus. Amplitudes are normalized to the maximum, and noise level PERG amplitudes values at birth were assumed. (A) Ages of half-maximal responses in control and Cop-1–treated rats were 4.4 and 7.4 weeks, respectively (F test, P < 0.0001). (B) Logistic curves of the falling phase did not differ between control and Cop-1–treated rats (F test, P = 0.66); therefore, data of both groups has been fitted into a single curve. Age of half-maximal response was 38.0 weeks.

Maturation dynamics of the highest spatial frequency tested (Fig. 3B, 0.368 cyc/deg) followed a slightly different course. Peak amplitude was reached earlier, at age 7 weeks (as opposed to 18 weeks in all other spatial frequencies), and this mature level of PERG amplitudes did not change significantly (ANOVA, P = 0.91) until age 26 weeks, after which significant attenuation was noted (Fig. 3B, weeks 26–59, ANOVA, P = 0.03). 

A gradual decline of PERG amplitudes began from age 26 weeks for the highest spatial frequency and from age 30 weeks for all other stimuli. The rate of decline was 7.2% to 8.2% per month in the four lowest spatial frequencies and 4.4% in the highest spatial frequency. By age 59 weeks, PERG amplitudes of all spatial frequencies declined back to immature levels of 5-week-old rats (Fig. 3B). Age50 of the declining amplitude phase was 38.0 weeks (Fig. 4B, r ² = 0.37). 

Cop-1–Immunized Rats Were Not Protected from Aging Changes but Had Altered Maturation Dynamics of PERG Amplitudes

Cop-1–treated rats had significantly lower N2 amplitudes than control animals at ages 5 and 59 weeks and had significantly higher amplitudes at ages 22 and 26 weeks (Fig. 5, Student's t-test, P < 0.05). 

Figure 5. 

 

Age-related changes in PERG N2 amplitudes of Cop-1–treated rats, compared with control rats. Mean responses (+SD) to a 0.046 cyc/deg stimulus are shown. Cop-1–treated rats were not protected from aging changes but had a significantly different course of PERG dynamics (ANOVA with repeated measures for within-group effects P = 0.001, Huynh-Feldt test). At the start and end points, Cop-1–treated rats had lower PERG amplitudes, while their maximal amplitudes were reached later (week 26) and were significantly higher. n = seven rats, 14 eyes in Cop-1 group, and n = 12 eyes of six rats in control group. Student's t-test, *P < 0.05.

No significant differences were found between amplitudes of Cop-1 and control rats when data from all time points pooled together were considered (ANOVA with repeated measures, between-group effects, P = 0.137). However, the dynamics of the maturation phase in Cop-1–treated rats differed from those of control animals, as evidenced by the significant age and group interaction (ANOVA with repeated measures for within-group effects, P = 0.001, Huynh-Feldt test). Furthermore, Age50 was reached later in Cop-1–treated rats compared with control rats, as was demonstrated by logistic curve fits of the 0.046 cyc/deg data. Namely, the Age50 values of Cop-1–treated and control rats were 7.4 and 4.4 weeks, respectively (Fig. 4A, F test, P < 0.011). Peak amplitudes were also reached later in Cop-1–treated rats compared with controls, namely, by the 26th week in response to the three lowest spatial frequencies and by the 22nd and 18th weeks in the two highest spatial frequencies (0.184 and 0.368 cyc/deg, respectively). In control rats, peak amplitude was reached by the 18th week of age (Fig. 5), except for responses to the highest spatial frequency, which peaked at age 7 weeks (Fig. 3B). 

The dynamics of the falling phase of Cop-1–treated rats were similar to those of the control rats. Logistic curves of the falling phase did not differ between control and Cop-1–treated rats (F test, P = 0.66), and, therefore, data of both groups has been fitted into a single curve (Fig. 4B). As in the control group, by age 59 weeks PERG amplitudes of Cop-1–treated rats also declined back to values recorded at age 5 weeks (Fig. 5, first and last time points). Therefore, we conclude that the PERG responses of Cop-1–immunized rats have extended maturation period but that they are not protected from aging changes of inner retinal function. 

Age-Related Implicit Time Changes in Control Rats

Unlike the PERG amplitudes, P1 implicit times did not change significantly during weeks 5 to 44 of life (ANOVA with repeated measures, within-subject effects, P > 0.08), with the exception of the 0.368 cyc/deg stimulus, where a trend for an age-dependent increase in implicit time was noticed (P = 0.02). Implicit times averaging 75 to 85 ms were measured for the other four spatial frequencies during this period (Fig. 6, data is shown for the 0.046 cyc/deg stimulus). From age 44 to 59 weeks, a significant decrease to 65 ms was noticed in all spatial frequencies (Student's t-test, P < 0.018). 

Figure 6. 

 

Age-related changes in average (+SD) PERG P1 implicit times of control rats, in response to 0.046 cyc/deg stimulus.

Age-related changes in N2 implicit times are shown in Figure 7. These changes were similar across the different spatial frequencies (data not shown), and therefore only data for the 0.046 cyc/deg stimulus is presented. These changes were biphasic. Implicit times of all spatial frequencies declined significantly from 140 to 150 ms at age 5 weeks and reached a plateau of approximately 130 ms by age 18 weeks (ANOVA with repeated measures, P < 0.0001). The second phase of decline was between 40 and 59 weeks, when implicit times decreased to 110 to 115 ms (ANOVA with repeated measures, P < 0.001). 

Figure 7. 

 

Age-related changes in average (+SD) PERG N2 implicit times of control rats, in response to 0.046 cyc/deg stimulus.

Implicit times of both P1 and N1 in Cop-1–treated rats followed similar courses as controls (ANOVA, P = 0.08 and 0.82, respectively). 

Discussion

In the present study, we characterized age-related changes of the Lewis rat PERG during the first 14 months of life. Most notably, these changes include a maturational period of increasing amplitudes spanning the first 7 weeks of life, followed by a mature period of approximately 23 weeks, and finally gradual attenuation, which we consider to be senescent in nature, starting at age 30 weeks, back to the amplitudes measured at age 5 weeks. 

Numerous factors can affect PERG amplitudes, including retinal, ³⁴ optical, ³⁵ visual stimuli, and anesthesia-related ones. We controlled for the nonretinal variables. Visual stimuli were kept constant throughout the experimental period. Pupils were dilated pharmacologically, and no anterior segment opacities were noted. Therefore, we consider the changes documented here to be mostly of retinal origin, more specifically reflecting age-related changes in inner retinal function. This is because the PERG signal reflects activity of RGC as was shown in studies of optic nerve transection in several species, including rats, ^13,36,37 as well as in intraretinal current recordings. ³⁸ More evidence was provided by pharmacological blockade of spiking activity with tetrodotoxin and glaucoma models. ³⁹ One potential limitation of our study is the unknown conductance across the eye globe, which might change with age and therefore affect voltage measurements. 

Although PERG amplitude can reflect the density of RGCs, ^10,21 the maturation (i.e., increasing amplitudes) period of the rat PERG probably overlaps a period of postnatal loss of redundant RGCs. ⁴⁰ Therefore, we assume that rather than reflecting an increase in RGC density, the PERG maturation period we demonstrated is a reflection of synapse maturation, which may not be completed before age 7 weeks. In mice, dendrites of several types of RGCs are properly restricted to the correct (i.e., adult) inner plexiform layer lamina by day 12 postpartum, ⁴¹ while other studies in mice report changes in RGC dendrite stratification to occur at least until 30 days postpartum. ^42,43 It is unknown when this process is completed in rats. Recently, Landi et al. report visual acuity, as measured by PERG in 5-week-old rats, to be enhanced as a result of insulin-like growth factor-1 treatment in days 1 to 7 postpartum. ⁴⁴ They find this effect to be mediated through brain-derived neurotrophic factor (BDNF) and tyrosine hydroxylase, which in turn act on dopaminergic amacrine cells. They also suggest a BDNF-mediated action on experience-dependent maturation of RGCs. ⁴² In our study, we documented a maturation process that lasts substantially longer than the 30-day postnatal period in mice. In view of our results, we believe that there is a need for studies of morphological synapse maturation in rats, similar to those conducted in mice. Furthermore, it should be remembered that there may be a lag between correct RGC dendrite localization and functional maturation of synapses, which could also account for the observed increase of PERG amplitudes till age 7 weeks. 

A developmental increase in PERG spatial resolution and decrease in P1 implicit time is documented in infants 5 to 20 weeks old. ³ We found the same trend in rats, although the implicit time reduction was a feature of the negative component of the PERG (Fig. 7). This difference might reflect the different generators of these two components. For example, in humans and macaque monkeys, optic nerve transection or degeneration and blocking of voltage-gated sodium channels (in macaques) preferentially attenuates N2 rather than P1, with the latter having shortened implicit times. ^40,45 In mice, Miura et al. ⁴⁶ find that, following optic nerve crush, amplitudes of both components are affected. Implicit times are not analyzed in that study, but based on the charts it seems that P1 is unaffected and N2 is shortened. 

The age at which senescent changes in human PERG are evident is hard to determine, as we are not aware of long-term, longitudinal studies in humans. By 72.3 (±7.5) years of age, a 40% reduction in PERG amplitudes, compared with 21.0 years (±6.0), is found. ⁷ Brecelj et al. find a significant, but weak, negative correlation to age in the PERG P50 amplitudes of schoolchildren aged 5 to 18 years. ⁸ The correlation between N95 amplitude (which is the analogue of N2 in the present study) and age is insignificant. In the present study, rats were individually tagged and repeatedly recorded throughout the first 14 months of life. This approach allowed us to determine the age at which aging changes take place. Assuming an expected life span of ∼2.35 years in Lewis rats, ⁴⁷ we found that the age-related decline in PERG amplitude begins after ∼25% of the rat's life span. This would correlate to approximately 20 years of age in humans, roughly equal to Brecelj's findings in P50. ⁸  

The most straightforward explanation for the low PERG amplitudes at the end point of this study is a decrease in RGC density. RGC density decrease over lifetime in the human retina is proposed to be 15% to 25% near the fovea, but is highly variable. ⁴⁸ Total RGC layer cell count estimated the rate of neural density decrease to be 0.29% and 0.53% per year in the macula and total retinal area, respectively. ⁴⁹ Consequently, a gradual decrease in retinal nerve fiber layer thickness is found. ⁵⁰ Similar age-related (proportional to life span) structural changes also occur in the RGC layer of rats and mice, with a density decrease rate of 1.5% and 2.3% per month, respectively. ²⁴ Significant RGC loss begins at age 9 to 12 months in mice (measured by total retinal fluorogold-positive counts) ⁵¹ and 4 to 12 months in rats (estimated by correlation of density decrease to retinal area expansion). ²⁵ While many authors attribute the decrease in RGC density to age-related RGC loss, Harman et al., ⁵² by means of sampling RGCs in rat retinal flatmounts and total retinal area measurements, finds no age-related RGC loss. Feng et al. ⁵³ reach the same conclusion, studying cross-sections of the entire rat retina. Katz et al. ²⁵ calculate the degree of RGC density loss in rats that cannot be explained by retinal stretching and attribute a loss of 10% at 12 months and 16% to 20% at 30 months to cell death. In the present study we found a decrease of ∼70% from peak amplitudes to the end point at age 59 weeks. This is a strikingly larger decrease than the 10% cell loss calculated by Katz et al. ²⁵ during a matching period. Therefore, we consider retinal stretching to be a significant factor in the PERG attenuation we found. Because the illuminated retinal area was kept constant throughout the study, lower RGC density can result in PERG amplitude reduction. 

One limitation of our study was that only inner retinal function was evaluated. As PERG is a downstream signal that depends on the integrity and input of photoreceptors and bipolar cells, the picture is not complete. Age-related changes in the outer and inner retinal function are studied by Charng et al. ⁵⁴ They compare FERG recordings in pigmented (Long-Evans) and albino (Sprague-Dawley) rats 3 and 18 months old and conclude that albino rats suffer from greater age-related attenuation of rod, ON-bipolar, and cone amplitudes. As PERG is mainly a light-adapted response, ^20,27 the latter might contribute to the low amplitudes we reported at age 59 weeks. However, our results are not in agreement with the age-related adaptation Charng et al. propose for the pSTR, the FERG parameter that reflects RGC activity. 

We chose to conduct this study using Lewis rats because this strain is commonly used in inner retinal research ^55–59 and, more specifically, it served successfully in PERG evaluation and Cop-1 immunization. ^21–23 The use of an inbred albino rat strain may be regarded as another potential limitation of this study, as some albino strains are reported to be susceptible to light-induced photoreceptor degeneration. ^60,61 However, early light- and age-related photoreceptor degeneration, which is well documented in albino Fischer 344 rats, ^62–65 is not reported in Lewis rats. Another widely used albino strain, the Sprague-Dawley, also appears to be less susceptible to degeneration, as its photoreceptor and outer nuclear layers were more preserved compared to Fischer 344 rats at 18 months of age. ⁶²  

There are conflicting reports on the degree of susceptibility of Lewis rats to excessive light-induced photoreceptor degeneration, with some researchers reporting increased damage ⁶⁶ while others demonstrate more preserved retinal structure. ⁶⁷ However, the rats in the present study were kept at low illumination levels (∼6 ft-c); therefore, they were less prone to light-induced damage, ⁶⁸ although this possibility cannot be ruled out. 

At 59 weeks of age, Lewis rats have reached approximately half their life span. ⁴⁷ In this context, the use of the word “senescence” might seem a misnomer, though it has been used to characterize age-related changes in rats 18 months (72 weeks) old. ⁵⁴ However, our goal was to propose an animal model for senescent changes, which may be defined as a gradual age-related deterioration in an apparently healthy animal. While these aging processes may be accelerated by strain-related factors, we believe that the model is still valid, as in the case of Fischer 344 rats, which serve as an accepted model for aging despite a short life span ⁶⁹ and photoreceptor degeneration. ⁶²  

In a previous study in our laboratory, rat PERG amplitudes were found to peak by age 11 weeks and decline thereafter. ²³ In the present study, we found that mature PERG levels (defined as 75% of maximal response) indeed develop by age 7 weeks, but that they peak later at age 18 weeks and, furthermore, begin aging changes only at age 26 to 30 weeks. These discrepancies could be related to the relatively large intersession variability of PERG recordings in rodents. For example, Porciatti et al. report fluctuations as high as 30% in successive PERG recordings in individual mice. ³⁴ These levels of variation might become significant in small sample studies, which is the case here. Alternatively, the differences between our current and previous studies could be due to the later end point in this study. Similar to Ben-Shlomo's study, ²³ we found an early peak (at 9 weeks) that was followed by two time points of lower average amplitude. But the prolonged duration of our study revealed the 9-week peak to be only temporary, followed by insignificantly fluctuating amplitudes spanning a total of 23 weeks (from weeks 7–30), until a significant and long-term decrease was documented. 

Age-related dynamics of PERG amplitudes in response to the highest spatial frequency differed from the other spatial frequencies mainly by peaking as early as age 7 weeks, compared with 18 weeks for the lower spatial frequencies. The highest spatial frequency tested was 0.368 cyc/deg, which is near the visual acuity limit of albino rats. ⁷⁰ Although a direct correlation of stimulus size to a specific RGC subtype is not established in the rat, Partridge et al. report a candidate RGC population. They find RGCs with a short and dense dendritic field (termed RGb2) that resembles feline β cells, which accounts for the higher visual acuity in the latter. ⁷¹ One may assume that the short dendritic tree allows for earlier completion of synaptic connections and, hence, faster maturation. 

Our second aim was to test whether the functional maturation and aging of the inner retina are affected by neuroprotective Cop-1 immunization. The higher peak amplitudes and the extended period of amplitude increase in Cop-1–immunized rats (Figs. 4, 5) might stem from neurotrophic factors secreted by specifically activated T-cells. Such an effect is found by Landi et al., who show that elevated BDNF levels at ages 1 to 7 days postnatal result in improved visual acuity at age 25 days. ⁴⁴  

Cop-1, being a weak agonist of myelin antigens that cross-reacts with myelin-based protein (MBP), can induce autoimmune neuroprotection, as is demonstrated in optic nerve crush, ²⁶ glaucoma, ²² and glutamate excitotoxicity ⁷² models. At the molecular level, Cop-1 immunization produces higher BDNF and lower NT-3 levels, compared with MBP immunization. ²⁶  

A key step in the abovementioned neuroprotection studies is chemotaxis of the activated T-cells to the injury site. We did not induce a cue for the T-cells to be attracted to the retina or the optic nerve, and, therefore, we cannot assume a preferential localization of T-cells to the inner retina of our rats. However, one could speculate about the possible presence of activated T-cells in nearby RGCs. Kipnis et al. inject Cop-1–activated T-cells to naïve rats; and, after 3 days, such cells are found in the optic nerve, although to a lesser extent than in injured optic nerves. ²⁶  

In conclusion, our 14-month-long study provides an accurate measure of age-related changes of inner retinal function in rats by means of repeated testing at the individual level. We found the functional development period to extend beyond the reported period of structural changes. This study also re-establishes the need for age-matched cohorts in PERG studies. 

Acknowledgments

The authors thank T. Bdolah-Abram for her assistance in statistical analyses. 

Footnotes

 Supported by the Joseph Alexander Foundation.

Footnotes

 Disclosure: S. Sandalon, None; R. Ofri, None